Research On Demand Forecasting Of Power Grid Material Based On Improved BP Neural Network

As the systems of "Intensive Management in human resources, financial resources, material resources and large-scale movements in programming, construction, operation, overhaul and production" are in progress in State Grid Corporation of China, the intensification of material management has received increased attention. The scope of centralized procurement of materials of Power Supply Companies becomes wider and wider, which means that yearly MRP provided by provincial electric companies will demand higher standards by State Grid Corp. But the existing yearly MRP tend to be made according to human experience, which lacks of scientific basis. So demand forecasting of power grid materials has practical implications. In addition, for supply management of the power grid material, demand forecasting also has important significance in reasonable arrangement of material bidding, improving the utilization of funds and reducing the backlog of materials.This paper analyzes the demand characteristics of power grid supplies and describes the characteristics of the project and non-project materials with the classification of the uses. Finally it determines the research mentality that demand forecasting of power grid materials should be conducted on the basis of single project.Meanwhile, this paper also analyses the existing prediction methods for the demand of materials and finally chooses the model of BP Neural Network to be used for the prediction of the demand of power grid materials. And then the paper describes the structure of BP neural network, learning preferences and learning algorithms in detail, and makes improvements on the training algorithm by SCG algorithm and initial weights and threshold selection by genetic algorithm on the basis of standard BP neural network. And this paper also supplies network design of the improved BP neural network prediction model, and finally makes case analysis with ACSR which is attached to the new110kV line construction materials to verify the validity and the science of the model, and the applicability of the model was discussed in detail.